Amit Malhotra

[Elon Musk @elonmusk]

This is exactly right. Source code is on the verge of becoming like assembly.

The next step is getting rid of “source code” entirely and just making an efficient binary directly with AI.

[Andrej Karpathy @karpathy]

Judging by my tl there is a growing gap in understanding of AI capability.

The first issue I think is around recency and tier of use. I think a lot of people tried the free tier of ChatGPT somewhere last year and allowed it to inform their views on AI a little too much. This is a group of reactions laughing at various quirks of the models, hallucinations, etc. Yes I also saw the viral videos of OpenAI's Advanced Voice mode fumbling simple queries like "should I drive or walk to the carwash". The thing is that these free and old/deprecated models don't reflect the capability in the latest round of state of the art agentic models of this year, especially OpenAI Codex and Claude Code.

But that brings me to the second issue. Even if people paid $200/month to use the state of the art models, a lot of the capabilities are relatively "peaky" in highly technical areas. Typical queries around search, writing, advice, etc. are *not* the domain that has made the most noticeable and dramatic strides in capability. Partly, this is due to the technical details of reinforcement learning and its use of verifiable rewards. But partly, it's also because these use cases are not sufficiently prioritized by the companies in their hillclimbing because they don't lead to as much $$$ value. The goldmines are elsewhere, and the focus comes along.

So that brings me to the second group of people, who *both* 1) pay for and use the state of the art frontier agentic models (OpenAI Codex / Claude Code) and 2) do so professionally in technical domains like programming, math and research. This group of people is subject to the highest amount of "AI Psychosis" because the recent improvements in these domains as of this year have been nothing short of staggering. When you hand a computer terminal to one of these models, you can now watch them melt programming problems that you'd normally expect to take days/weeks of work. It's this second group of people that assigns a much greater gravity to the capabilities, their slope, and various cyber-related repercussions.

TLDR the people in these two groups are speaking past each other. It really is simultaneously the case that OpenAI's free and I think slightly orphaned (?) "Advanced Voice Mode" will fumble the dumbest questions in your Instagram's reels and *at the same time*, OpenAI's highest-tier and paid Codex model will go off for 1 hour to coherently restructure an entire code base, or find and exploit vulnerabilities in computer systems. This part really works and has made dramatic strides because 2 properties: 1) these domains offer explicit reward functions that are verifiable meaning they are easily amenable to reinforcement learning training (e.g. unit tests passed yes or no, in contrast to writing, which is much harder to explicitly judge), but also 2) they are a lot more valuable in b2b settings, meaning that the biggest fraction of the team is focused on improving them. So here we are.

[Greg Brockman @gdb]

Software development is undergoing a renaissance in front of our eyes.

If you haven't used the tools recently, you likely are underestimating what you're missing. Since December, there's been a step function improvement in what tools like Codex can do. Some great engineers at OpenAI yesterday told me that their job has fundamentally changed since December. Prior to then, they could use Codex for unit tests; now it writes essentially all the code and does a great deal of their operations and debugging. Not everyone has yet made that leap, but it's usually because of factors besides the capability of the model.

Every company faces the same opportunity now, and navigating it well — just like with cloud computing or the Internet — requires careful thought. This post shares how OpenAI is currently approaching retooling our teams towards agentic software development. We're still learning and iterating, but here's how we're thinking about it right now:

As a first step, by March 31st, we're aiming that:

(1) For any technical task, the tool of first resort for humans is interacting with an agent rather than using an editor or terminal.
(2) The default way humans utilize agents is explicitly evaluated as safe, but also productive enough that most workflows do not need additional permissions.

In order to get there, here's what we recommended to the team a few weeks ago:

1. Take the time to try out the tools. The tools do sell themselves — many people have had amazing experiences with 5.2 in Codex, after having churned from codex web a few months ago. But many people are also so busy they haven't had a chance to try Codex yet or got stuck thinking "is there any way it could do X" rather than just trying.
- Designate an "agents captain" for your team — the primary person responsible for thinking about how agents can be brought into the teams' workflow.
- Share experiences or questions in a few designated internal channels
- Take a day for a company-wide Codex hackathon

2. Create skills and AGENTS[.md].
- Create and maintain an AGENTS[.md] for any project you work on; update the AGENTS[.md] whenever the agent does something wrong or struggles with a task.
- Write skills for anything that you get Codex to do, and commit it to the skills directory in a shared repository

3. Inventory and make accessible any internal tools.
- Maintain a list of tools that your team relies on, and make sure someone takes point on making it agent-accessible (such as via a CLI or MCP server).

4. Structure codebases to be agent-first. With the models changing so fast, this is still somewhat untrodden ground, and will require some exploration.
- Write tests which are quick to run, and create high-quality interfaces between components.

5. Say no to slop. Managing AI generated code at scale is an emerging problem, and will require new processes and conventions to keep code quality high
- Ensure that some human is accountable for any code that gets merged. As a code reviewer, maintain at least the same bar as you would for human-written code, and make sure the author understands what they're submitting.

6. Work on basic infra. There's a lot of room for everyone to build basic infrastructure, which can be guided by internal user feedback. The core tools are getting a lot better and more usable, but there's a lot of infrastructure that currently go around the tools, such as observability, tracking not just the committed code but the agent trajectories that led to them, and central management of the tools that agents are able to use.

Overall, adopting tools like Codex is not just a technical but also a deep cultural change, with a lot of downstream implications to figure out. We encourage every manager to drive this with their team, and to think through other action items — for example, per item 5 above, what else can prevent a lot of "functionally-correct but poorly-maintainable code" from creeping into codebases.

[Sridhar Vembu @svembu]

What is ailing the software job market is not AI taking away jobs (not yet anyway).

Here is my thesis, as a participant and observer of software for 30 years. Over those 30 years:

1. Massive over-capacity steadily developed in enterprise software due to a flood of VC, PE and IPO money.

2. Software vendors applied liberal doses of marketing spending to spread Fear (of missing out) and Uncertainty ("tech is changing, you need us") and Doubt ("are you confused? trust us") among corporate customers and the result was ever growing IT spending. Enterprise IT budgets kept rising because which CIO or Board would want to be seen as lagging? Major corporations in the West have layers and layers of duplicated IT systems, lots of money spent to acquire them and even more money spent to get the disparate systems to work together. The more inefficient IT systems ended up becoming permanent resource drains, requiring vast human resources to simply keep them running.

3. Large enterprises in the West found those needed human resources in India - transferring those inefficiencies to IT services firms in India, often multiplied by 3 or even 4! That "multiplied inefficiency" happened because IT budgets were kept fixed in dollar terms and more people got hired in India to "get more done".

Result? A large number of IT jobs in India came to depend on those original inefficiencies and the multiplied inefficiencies.

4. Note how Indian banks and financial institutions are more IT savvy, while spending far less, than established financial majors in the US - Indian firms did not have the budgets to splurge! Necessity made them highly efficient and today India's financial institutions are able to fend off foreign competition easily.

5. One of the deep facts about software is that often a 2 person team can solidly outperform a 20 person team and a 10 person team can do the work of a 200 person team. This is not just due to talent disparity - even when the large teams have equivalent talented people, they can easily end up being wasted on unproductive projects.

6. When people are billed by the hour or by the staff month (input metrics), as it happens often in IT services, there is zero incentive to remove those inefficiencies, to empower the 2 person team as opposed to hiring 20.

It is those multiplied inefficiencies in IT, built up over decades, that is facing a reckoning now.

Now enter AI. A large amount of code is boilerplate code in many projects and AI can eat such code for breakfast. Depending on the nature of a project AI can offer 10-20% productivity gains. Significant but not the 10x or 100x leap yet to destroy jobs on a vast scale.

But those AI gains of today pale in comparison to the "multiplied inefficiencies" built up over decades.

This reckoning nearly happened in the aftermath of the Global Financial Crisis (GFC 2008-9). It got postponed because the central banks flooded the world with money and a part of that money flooded into enterprise IT.

The pandemic unleashed another flood of money into enterprise IT.

Those floods are now history and we have a serious drought.

That is why I am pessimistic about the software job market, even before accounting for AI.

[Amit Malhotra @amitx]

Join in for real world stories on using AI to super charge your business. https://t.co/9ouK2p9Hev

[Shakeel Rashed @Shakeelr]

Multidimensional data exists in many fields & is a difficult problem to solve. 😢our efforts at @PatternSciences did not work in the target market - Mathematician uncovers methods to shrink sampling errors in large-dimensional data sets https://t.co/Tb7BbE41kn via @physorg_com

[Amit Malhotra @amitx]

@beckershr Mark Cuban's pharmacy started with a cold email https://t.co/phZ7Hh9SXX
"..if this guy can jack up the prices 750% for lifesaving medicines, can we go the opposite .? ..Are there inefficiencies in this industry that really allow us to do it and really make a diff..?'"

[Elon Musk @elonmusk]

The duty of a leader is to serve their people, not for the people to serve them

[Amit Malhotra @amitx]

@elonmusk Help me avoid marital strife. When are the Austin made Model Ys with the new cells going to be delivered??? @ModelYNews
https://t.co/NZnqX7vpBv

[Meghan Bobrowsky @MeghanBobrowsky]

For @amitx and his wife, the decision to sell their 2018 Model 3 named Carrol happened fast and unexpectedly.

Amit went out to get the bumper fixed. Then Radhika received a call.

“This is one of your quintessential he went out to buy bread and sold my Tesla” stories, she said.

[Meghan Bobrowsky @MeghanBobrowsky]

For @amitx and his wife, the decision to sell their 2018 Model 3 named Carrol happened fast and unexpectedly.

Amit went out to get the bumper fixed. Then Radhika received a call.

“This is one of your quintessential he went out to buy bread and sold my Tesla” stories, she said.

[Amit Malhotra @amitx]

@Shakeelr @JoshuaBaer @JoshuaBaer @Shakeelr Guess what came out of this conversation!
https://t.co/FmUBQ3Q5gw

https://t.co/Na2JbSbZCu

[Meghan Bobrowsky @MeghanBobrowsky]

For @amitx and his wife, the decision to sell their 2018 Model 3 named Carrol happened fast and unexpectedly.

Amit went out to get the bumper fixed. Then Radhika received a call.

“This is one of your quintessential he went out to buy bread and sold my Tesla” stories, she said.

[Amit Malhotra @amitx]

@bsierakowski @JoshuaBaer regular suspects locally Henna Chevrolet, Carmax, the Sewell/Land Rover, any Autonation dealer. Online auctions, try carandbids and bringatrailer for the best price (detail and take pics). And then you have Vroom, caravna shift and others who should give you a quote with a Vin.

[Shakeel Rashed @Shakeelr]

Would love a conversation between @nntaleb @yudapearl ...maybe you can start off with a mutual topic of Kahneman and his work since both of you refer to him and one of my favorite intellectuals

[Pattern Sciences @PatternSciences]

About to talk Real World Evidence at @UTAustin

[Shakeel Rashed @Shakeelr]

When the book is so good, that you wake up an hour earlier to read a few chapters before the day starts…love @WalterIsaacson ‘s writing, and how human and relatable he makes every person mentioned

[Pattern Sciences @PatternSciences]

We are in the right business #clinicaltrials adding @CTTI_Trials

[Andy Mychkovsky @healthcareandy]

*Investor at JPM*

- Digital health is overrated.
- Founders overvalue their ideas.
- Genomics is still far off.
- Glad it’s not raining.
- Value-based care doesn’t work.
- Doctors don’t have SAAS margins.
- Where’s the open bar?

Miss anything?

#JPMorgan2020 #JPM2020 #JPMHC20

[Amit Malhotra @amitx]

Are we rational? Kahneman - Judgement, heuristics, biases, and action by @Shakeelr https://t.co/tKhGHoidVf

[Amit Malhotra @amitx]

Thank you! Super excited to be working with @CFDefense using AI to train the next generation of warfighters.

[Fed Supernova @FedSupernova]

A huge congratulations to https://t.co/QkMvNrP3W5 for winning the 100K Challenge! Excited to see your future endeavors

[Fed Supernova @FedSupernova]

Now we have https://t.co/qIlH8s8OqR Pattern Sciences is a professional services company and research lab that brings a pragmatic approach towards AI and machine learning in the complex domains.